SE1150722A1 - Demodulation of OFDM QAM signals with channel estimation error - Google Patents

Abstract

The present invention relates to a method, wherein channel estimation errors observed in unequal amplitude constellations in the additive noise and frequency equalization process in OFDM (Orthogonal frequency-division multiplexing) based multi-carrier systems are measured by means of a receiver and a transmitter, new constellations are generated according to the said measurements, and the errors are compensated.

Description

the frequency domain is often used in OFDM-based systems because they are easy to implement. Channel taps are estimated based on the transmission of a known data sequence.

This procedure is called data-guided or pilot-based channel estimation.

The baseband model for a transmitted OFDM symbol can be described as: äXnejZIIicn / N 3 xk = f fi2fi g ..., 1v-1 (1) where N represents the FFT size, n indicates index in the frequency domain, k indicates index in the time domain and {Xn} indicates the transmitted symbol / bit sequence.

The received OFDM symbol at the output of the FFT block can be denoted as: -j 21rkn / N Rn = i N-1 mk = 0 Tke n = 0.1, ..., 1v-1 (2) rk in equation 2 is the received symbol sample.

The received OFDM symbol at the output of the F FT block can also be expressed as: Rn = HnXn + In + Wn (3) In Equation 3, Hn represents the nth channel pin, ln represents the interchannel interference on the nth channel and Wn the additive white Gaussian noise component (AWGN, "Additive White Gaussian Noise").

Channel estimation is performed using the known pilot symbols.

The channel estimate can mainly be obtained by the following equation, _ Rni H ”_ Ti <4) In equation 4, P denotes the izte pilot symbol.

Where the number of pilot symbols used in the system is L, the channel pins used in the FDE can be obtained by averaging the channel estimates (5) per symbol as follows: ^ -l .L ^.

Hn _ 'LZt = 1Hn, t (5) The samples received at the output of F DE are 10 15 20 25 30 _ = gfx "+ få +: Ti (6) As can be seen from the explanations given above, the estimation error associated with channel loss estimate, Ûn, performance deterioration in OF DM system.

As a result of the experimental and analytical studies that have been performed, it is observed in the prior art applications that in the OFDM-based communication methods, the noise applied to the transmitted data and the interchannel interference (ICI) occur more in constellation points with high amplitudes than the points close to the origin, where the real (in-phase) and virtual (quadrature) amplitudes of the quadrature axis are zero (fi gur 4 and 5).

Hn = En + AH "(7) In Equation 7, AH" represents the channel estimation error. The received symbol can then be expressed as zn = xn To demonstrate this effect, a 64-QAM constellation is used.

A common 64-QAM constellation is shown in Figure 4. The average symbol energy in the constellation is normalized to 1. Samples received from this constellation through ideal channels with 3 pilot symbols are shown in Figure 5. According to what can be observed from this Figure, constellation points with higher amplitude values exposed to more noise during channel estimation.

The solution methods provided in the prior art for the problem of improving the accuracy of channel estimation are generally based on the transmission of a known data sequence called pilot symbols.

One of the applications in the prior art is described in US patent document US-63273 14. The said document is intended to improve the channel estimation method for a more accurate channel frequency response.

Another application in the prior art is described in US patent document US-2004240376. In the mentioned document, a virtual learning symbol is created and processed to correct the channel estimation error. 10 15 20 25 30 35 Summary of the Invention The object of the present invention is to provide a method in which OFDM-based receivers and transmitters are used, in which the deterioration in performance which arises due to channel estimation errors is reduced.

Another object of the invention is to provide a method utilizing an OFDM-based receiver, in which symbol errors are reduced by utilizing the various amplitude constellation points, and a transmitter operating in accordance with said receiver.

A further object of the invention is to provide a method which utilizes an OFDM-based receiver, in which the error levels of the high amplitude signals are reduced, and a transmitter.

Detailed Description of the Invention The method achieved to achieve the object of the present invention is illustrated in the accompanying figures, in which Figure 1 shows the block diagram of a transmitter according to the prior art.

Figure 2 shows the block diagram of an embodiment of the receiver according to the present invention.

Figure 3 shows the block diagram of the transmitter operating in accordance with the receiver according to the present invention.

Figure 4 is the diagram showing the arrangement of the points in a conventional 64-QAM constellation according to the prior art.

Figure 5 is the diagram showing the arrangement of the points in a conventional 64-QAM constellation according to the prior art, when noise is added thereto.

Figure 6 is the fate diagram of the calculation of the noise ratio in the invention.

Figure 7 is the fate diagram showing the operation of the constellation generator.

Figure 8 is the diagram showing the arrangement of the points in 64-QAM constellation points distorted by the invention.

Figure 9 is the diagram showing the error levels of the 64-QAM constellation points distorted with the invention and the 64-QAM constellation points according to the prior art.

The parts of the figures have all been given a reference numeral, the designations refer to the following: 1. Receiver 2. Analog-to-digital converter 10 15 20 25 30 3. Cyclic pre fi x / suf fi x remover 4. Serial-to-parallel converter 5. N-FFT block 6. Channel estimator 7. FEQ equalizer for frequency domain 8. Demodulator 9. Serial-to-parallel converter 10. Fault control decoder 1 l. Channel error processing unit 12. Constellation generator 21. Transmitter 22. Fault control encoder 23. Modulator 24. Serial to-parallel-converter 25. N-IFFT-block 26. Parallel-to-serial1-converter 27. Cyclic prefix / suf fi x-adder 28. Digital-to-analog converter The receiver (1) according to the invention comprises an analog-to-analog digital converter (2), a cyclic pre fi x / suf fi x remover (3), a serial-to-parallel converter (4), an N-FFT block (5), a channel estimator (6), an equalizer (7 ) for frequency domain, a demodulator (8), a serial-to-parallel converter (9), an error control decoder (10) and a channel error processor ignition unit (1 1).

The transmitter (21) operating in accordance with the receiver (1) according to the invention comprises an error check encoder (22), a modulator (23), a serial-to-parallel converter (24), an N-IFFT block (25). ), a parallel-to-serial converter (26), a cyclic prefix / suffix adder and a digital-to-analog converter (28).

Signals with higher amplitude values are exposed to more additive noise, and this invention is realized for this reason for different amplitude constellations.

The performance degradation observed in high amplitude symbols due to the fact that more noise occurs therein is corrected in the present invention by using a feedback channel between the receiver (1) and the transmitter (21).

The noise ratio in the receiver (1) and the transmitter (21) is measured by the following steps: - the pilot symbol is transmitted by means of the transmitter (21) (101), 10 15 20 25 30 - the pilot symbol is received by the receiver (1) (102), - the demodulator (8) calculates the observed noise ratio (103), - the channel error processing unit (11) calculates the ratio between the symbol amplitude and the observed noise ratio (104), - the channel error processing unit (11) calculates the expected noise ratio of the non-transmitted symbols (105).

A demodulator (8) is used in the invention to reduce the symbol error. In the demodulator (8) the decision limits are destroyed. How much noise is expected in which symbol is determined by calculating the noise ratio. The decision limits are distorted according to the expected noise in the demodulator (8) in the receiver (1).

The receiver (1) in an embodiment of the invention comprises a constellation generator (12). In this embodiment, the receiver (1) determines the most suitable constellation by means of the constellation generator (12). In this embodiment, the decision limits and the constellation to be used are determined by the following steps, performed by the constellation generator (12): - adapting the calculated noise ratio to the constellation used (201), - calculating the distances between constellation points to be used according to the noise ratios (202), - generate a new constellation (203), - check whether the same error ratio is given with the generated constellation (204), - if the same error ratio is provided with the generated constellation, ensure that the constellation is used in the transmitter (21) and corresponding decision limits are used in the receiver (1) (205), - if the same error condition is not provided with the generated constellation, return to step 203 and continue to generate new constellations.

In another embodiment of the invention, the performance degradation observed in high amplitude symbols due to the fact that more noise takes place therein is determined by testing specific constellations using a feedback channel between the receiver (1) and the transmitter (21). The constellation observed to have minimal error ratio is selected for use in the demodulator (8) of the receiver (1). In an alternative embodiment, the constellation that is to provide the maximum performance is distorted by using a linear model. Performance tests are performed by testing various values of the parameters used, and the constellation with the maximum performance is selected for use in the system. In this solution, the distorted constellation is located in the receiver (1), while the distorted decision limits are located in the modulator (23) in the transmitter (21).

An exemplary embodiment of the invention is realized for a standard 64-QAM constellation ("Quadrature Amplitude Modulation"). In this embodiment, the constellation decision boundaries are distorted so that they will be different.

The decision limits are distorted in the demodulator (8). The decision limits are calculated by using the principle with the same Euclidean distance. In the demodulator (8), the constellation points with higher amplitudes are separated from the rest of the constellation points.

As a result of this alternative of the invention, the distorted constellation provides a gain of 0.5 dB in a non-attenuation channel compared to the usual constellation (fi Figure 9). The distortion in this figure is achieved in line with the observed additive error rate. The amplitudes of the symbols with the highest amplitudes increase further. This is because more additive noise is observed in these symbols due to the channel estimation error. In order for this noise not to give rise to symbol errors, high amplitude symbols should be separated as much as possible from the symbols close to them. This can be achieved by increasing the distance between the high amplitude symbols.

Distortion of the constellation is decided on according to the amount of error to be added to the transmitted data (fi gur 6). The error distribution factor to be applied to said data depends on the medium through which said data is to be transmitted.

It is possible to develop a wide range of embodiments of the method according to the invention. The invention may not be limited to the examples described herein and is substantially in accordance with the claims.

Claims (4)

10 15 20 25 30 PATENT REQUIREMENTS A method used in OFDM communication, which utilizes a receiver (1) comprising an analog-to-digital converter (2), a cyclic pre-x / suf-x remote (3), a serial-to-parallel converter ( 4), an N-FFT block (5), a channel estimator (6), a frequency domain equalizer (7), a demodulator (8), a serial-to-parallel converter (9), an error check decoder (10) and a channel error processing unit (1 1); and a transmitter (2) comprising an error check encoder (22), a modulator (23), a serial-to-parallel converter (24), an N-IFFT block (25), a parallel-to-serial1 converter (26), a cyclic preñx / suf fi x adder (27) and a digital-to-analog converter (28); characterized in that the measurement of the noise ratio is performed by the following steps: - the pilot symbol is transmitted by means of the transmitter (21) (101), - the pilot symbol is received by the receiver (1) (102), - the demodulator (8) calculates the observed noise ratio (103 ), - the channel error processing unit (11) takes the ratio between the symbol amplitude and the observed noise ratio (104), - the channel error processing unit (11) calculates the expected noise ratio of the non-transmitted symbols (105). Method according to claim 1, characterized in that the decision limits, in order to reduce the symbol error, are distorted in the demodulator (8) in line with the expected noise. A method according to claim 1 or claim 2, wherein a constellation generator (12) is used in the receiver (1); and characterized in that the constellation generator (12) performs the following steps: - adapting the calculated noise ratio to the constellation used (201), - calculating the distances between constellation points to be used according to the noise ratios (202), - generating a new constellation (203), - check whether the same error ratio is provided with the generated constellation (204), 10 - if the same error ratio is provided with the generated constellation, ensure that the constellation is used in the transmitter (21) and the current decision limits are used in the receiver (1) (205 ), - if the same error condition is not provided with the generated constellation, return to step (203) and continue generating new constellations. Method according to claim 1 or claim 2, wherein, by testing specific constellations using a feedback channel between the receiver (1) and the transmitter (21), the constellations observed to have the minimum error ratio are selected for use in the demodulator. (8) in the receiver (1).